CN106327738B - A kind of intelligent hierarchical monitoring system - Google Patents
A kind of intelligent hierarchical monitoring system Download PDFInfo
- Publication number
- CN106327738B CN106327738B CN201610744200.7A CN201610744200A CN106327738B CN 106327738 B CN106327738 B CN 106327738B CN 201610744200 A CN201610744200 A CN 201610744200A CN 106327738 B CN106327738 B CN 106327738B
- Authority
- CN
- China
- Prior art keywords
- module
- image
- information
- video
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B25/00—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems
- G08B25/01—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems characterised by the transmission medium
- G08B25/10—Alarm systems in which the location of the alarm condition is signalled to a central station, e.g. fire or police telegraphic systems characterised by the transmission medium using wireless transmission systems
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The intelligent monitor system of the present invention includes master controller, detection means, automatic alarm system and monitor terminal, the detection means includes the first detection means, second detection device and the 3rd detection means, and the master controller is used to realize the control and communication to first detection means, second detection device, the 3rd detection means, automatic alarm system.By the intelligent monitor system of the present invention, hierarchical monitoring is realized, target is accurately identified, quickly judges anomalous event, improves alarm accuracy rate, there is positive beneficial effect.
Description
Technical field
The present invention relates to information monitoring field, more particularly to a kind of intelligent hierarchical monitoring system.
Background technology
The life and living environment of safety increasingly cause the concern of people, and traditional video surveillance system can only provide monitoring
The image in region, the monitoring to monitor area, the identification to destination object, analysis, discrimination, tracking etc. fully relies on people to enter
OK.Therefore, traditional video surveillance needs related personnel to carry out continual monitoring.Intelligent video monitoring possesses automatic, intelligence
Image analysis capabilities,, can be with by automatically analyzing for the image that is shot to video camera in the case where not needing human intervention
Identification, dynamic object in scene is differentiated, obtain the information such as the size of target, quantity, direction, speed, can be in abnormal conditions
Alarm, video recording, tracking etc. are made when generation in a manner of most fast and optimal to react.
But intelligent monitor system, because the factors such as identification inaccuracy can cause to report by mistake and fail to report, it is unnecessary to cause on the contrary
Interference, it is to be badly in need of solution therefore, to improve detection, the accuracy for identifying and judging, reduce wrong report and failing to report phenomenon to greatest extent
Technical problem.
The content of the invention
The purpose of the present invention is achieved through the following technical solutions.
According to the embodiment of the present invention, a kind of intelligent hierarchical monitoring system is proposed, the system includes master controller, inspection
Device, automatic alarm system and monitor terminal are surveyed, wherein,
The detection means includes the first detection means, second detection device and the 3rd detection means, the master controller
For realizing to first detection means, second detection device, the 3rd detection means, the control of automatic alarm system and logical
Letter；
First detection means is used to carry out primary detection, judges to lead to whether indoor each entrance has exception
Activity, first detection means is set persistently to be detected if abnormal movement is not detected, otherwise, automatic alarm system
Into treating alarm status, and start second detection device and detected；
The second detection device is for detecting whether be anomalous event, if anomalous event, then automatic alarm system
Alerted, otherwise, start the 3rd detection means and detected；
3rd detection means be used to detecting it is indoor whether someone's activity, if nobody is movable, repeated priming the
One detection means is detected, if someone is movable, automatic alarm system enters no alarm state；
Automatic alarm system includes voice module and Cloud Server, when automatic alarm system enters alarm status, master control
Device control voice module processed sends voice warning, while warning message is sent into monitor terminal, Ke Yishe by Cloud Server
Put a variety of alarm sounds；
Monitor terminal is connected with second, third detection means and automatic alarm system by network respectively, for user
Monitoring situation is checked, monitor terminal receives video information and the identification of warning message and the transmission of second, third detection means
Object information, and it is determined as that wrong video feature information is sent into Cloud Server through terminal by what second detection device was sent
It is trained；
Cloud Server is connected with second detection device and monitor terminal by network respectively, for video frequency feature data
Training, according to the information of user feedback, dynamically updates the data sorting technique, wherein,
The video segment that Cloud Server is sent according to monitor terminal extracts characteristic information, and by characteristic information, classification results
Storage parses, and according to renewal sample training data, generate new as new sample to the instruction that monitor terminal is sent
Data classification method configuration file, while send update notification to second detection device.
According to an embodiment of the invention, the first detection means includes detecting module, each for testing staff's disengaging
The activity of individual entrance, door status sensor, infrared detector, photo-electric blocking can be used installed in the first detection means of entry door
Inductor, microwave sensor or dual technology detector, for detecting the personnel activity of disengaging entry door；Going out beside window
Entrance detection means, double curtain infrared detectors can be used, can be identified according to the triggering priority time of its two infrared probes
Personnel enter outgoing direction, tell personnel and are stretched out one's hand from the abnormal movement got in outside window, daily routines indoors and from interior
Close window activity；The first detection means of installation on balcony, can be used infrared detector or dual technology detector, for detecting personnel
In the activity of balcony.
According to an embodiment of the invention, second detection device includes：Information collecting device and anomalous event judge
Device, the anomalous event judgment means include the first transceiver module, analysis module, data categorization module, memory module and
First control module, control module are connected with modules respectively, wherein,
Information collecting device is camera and sound input system；
Analysis module is used to analyze and process the video data of information collecting device collection；
The image information that data categorization module is handled according to analysis module judges whether monitors environment is abnormal, the data point
Original data classification method configuration information is stored with generic module, data categorization module identifies which kind of current video belongs to
Afterwards, recognition result is sent to the monitor terminal of user, by the continuous feedback data of user, Cloud Server training generates new number
The configuration file being used for according to sorting technique, methods described in the data categorization module of renewal；
Memory module is used for the video data for storing collection, is easy to checking and playing back for user；
Transceiver module is used between anomalous event judgment means and information collecting device, monitor terminal and automatic alarm system
Information exchange；
First control module is the corn module of the anomalous event judgment means, respectively between control modules
Data communication and interaction.
According to an embodiment of the invention, the analysis module includes Target Acquisition module, the decomposition being linked in sequence
Module and feature extraction module, wherein,
The acquisition module, for using corresponding algorithm, the object of background and motion is distinguished, then extraction detects
Target；
The decomposing module, for after the acquisition module extracts the target detected, splitting to target；
The feature extraction module, for by being tracked to the target of segmentation, the characteristic information of image being extracted, by institute
Image feature information is stated to be preserved in the form of characteristic vector.
According to an embodiment of the invention, first transceiver module includes video access unit, data interaction list
Member, wherein video access unit are used for the video for accessing collection, and video is sent into analysis module and memory module, its
In,
Video access unit is video decoding circuit, and the video data for information collecting device to be gathered is decoded；
The data interaction that data interaction unit is used between anomalous event judgment means and monitor terminal, sorts data into mould
The classification results of block are sent to monitor terminal, while the video for receiving monitor terminal transmission checks information；It is also used for anomalous event
Information exchange between judgment means and Cloud Server, Cloud Server are sentenced by training video characteristic, notice anomalous event
Disconnected device updates the data the classification configurations file in sort module, and data interaction unit uses wired network communication or wireless network
Any one in communication mode, cable network can use Ethernet interface, and wireless network communication can be WIFI and 3G/
4G。
According to an embodiment of the invention, the second detection device also includes：Pattern recognition device, described image
Identification device includes：Processing module, image collection module, matching module, database, identification module, the second transceiver module composition,
Wherein,
Processing module, for the audio/video information of processing information acquisition module collection, extract the image for face recognition；
Image collection module, for obtaining the image information for face recognition, the digital information of described image is carried out
Analysis and processing, then carry out feature extraction, and the facial image information in digital information is obtained with mode identification method；
Matching module, whether the facial image information for judging acquired has storage in database, for being obtained
Facial image information, the content stored in searching data storehouse, be determined as when finding matched information without exception
Situation, send and start door opening command, sentenced when the content matched with the facial information is not present in database by anomalous event
Disconnected device judges whether it is anomalous event；
Database, for storing face information and the audio-frequency information that user pre-sets；
Identification module, for carrying out match cognization, the identification module bag according to existing face information and audio-frequency information
Include face-image identification and acoustic recognition unit；
Second transceiver module, for realizing second detection device and the communication of monitor terminal and automatic alarm system.
According to an embodiment of the invention, the face-image recognition unit obtains to target image initial data
Take, pre-process and using Invariance feature extracting method and neural network recognization technology；Acoustic recognition unit is directed to voice
Data signal is all to carry out match cognization with corresponding processing method and discrimination method, and the match is successful if having the target information,
Otherwise judge whether it is anomalous event by anomalous event judgment means.
According to an embodiment of the invention, the 3rd detection means includes：Video information acquisition module, detection module,
Control module and particular image monitoring module, wherein,
Video information acquisition module is used to gather video image, and the video image currently collected is sent to detection mould
Block；
Detection module is used to receive the input picture from video information acquisition module, and the moving image for exporting acquisition is extremely controlled
Molding block；
Control module is used to receive the moving image for carrying out self-detection module, and video prison is carried out according to the moving image received
Control；
Particular image monitoring module is used to, by merging night interesting target and background image between daytime, obtain final night
Activity detection image.
The intelligent hierarchical monitoring system of the present invention includes master controller, detection means, automatic alarm system and monitoring eventually
End, the detection means include the first detection means, second detection device and the 3rd detection means, and the master controller is used for real
Now to the control and communication of first detection means, second detection device, the 3rd detection means, automatic alarm system.Pass through
The intelligent monitor system of the present invention, realizes hierarchical monitoring, accurately identifies target, quickly judges anomalous event, improve alarm
Accuracy rate, there is positive beneficial effect.
Brief description of the drawings
By reading the detailed description of hereafter preferred embodiment, it is various other the advantages of and benefit it is common for this area
Technical staff will be clear understanding.Accompanying drawing is only used for showing the purpose of preferred embodiment, and is not considered as to the present invention
Limitation.And in whole accompanying drawing, identical part is denoted by the same reference numerals.In the accompanying drawings：
Accompanying drawing 1 shows the intelligent monitor system overall structure diagram according to embodiment of the present invention；
Accompanying drawing 2 shows the second detection device structural representation according to embodiment of the present invention；
Accompanying drawing 3 shows the anomalous event judgment means structural representation according to embodiment of the present invention；
Accompanying drawing 4 shows the analysis module structural representation according to embodiment of the present invention；
Accompanying drawing 5 shows the first transceiver module structural representation according to embodiment of the present invention；
Accompanying drawing 6 shows the pattern recognition device structural representation according to embodiment of the present invention；
Accompanying drawing 7 shows the 3rd structure of the detecting device schematic diagram according to embodiment of the present invention；
Accompanying drawing 8 shows the detection module structural representation according to embodiment of the present invention；
Accompanying drawing 9 shows the motion detection block structural representation according to embodiment of the present invention；
Accompanying drawing 10 shows the difference block structural representation according to embodiment of the present invention；
Accompanying drawing 11 shows the particular image monitoring module structural representation according to embodiment of the present invention.
Embodiment
The illustrative embodiments of the disclosure are more fully described below with reference to accompanying drawings.Although this public affairs is shown in accompanying drawing
The illustrative embodiments opened, it being understood, however, that may be realized in various forms the disclosure without the reality that should be illustrated here
The mode of applying is limited.Conversely, there is provided these embodiments are to be able to be best understood from the disclosure, and can be by this public affairs
The scope opened completely is communicated to those skilled in the art.
According to the embodiment of the present invention, a kind of intelligent hierarchical monitoring system, as shown in Figure 1, the system bag are proposed
Master controller, detection means, automatic alarm system and monitor terminal are included, wherein,
The detection means includes the first detection means, second detection device and the 3rd detection means, the master controller
For realizing to first detection means, second detection device, the 3rd detection means, the control of automatic alarm system and logical
Letter；
First detection means is used to carry out primary detection, judges to lead to whether indoor each entrance has exception
Activity, first detection means is set persistently to be detected if abnormal movement is not detected, otherwise, automatic alarm system
Into treating alarm status, and start second detection device and detected；
The second detection device is for detecting whether be anomalous event, if anomalous event, then automatic alarm system
Alerted, otherwise, start the 3rd detection means and detected；
3rd detection means be used to detecting it is indoor whether someone's activity, if nobody is movable, repeated priming the
One detection means is detected, if someone is movable, automatic alarm system enters no alarm state；
Automatic alarm system includes voice module and Cloud Server, when automatic alarm system enters alarm status, master control
Device control voice module processed sends voice warning, while warning message is sent into monitor terminal, Ke Yishe by Cloud Server
Put a variety of alarm sounds；
Monitor terminal is connected with second, third detection means and automatic alarm system by network respectively, for user
Monitoring situation is checked, monitor terminal receives video information and the identification of warning message and the transmission of second, third detection means
Object information, and it is determined as that wrong video feature information is sent into Cloud Server through terminal by what second detection device was sent
It is trained；Monitor terminal in the present invention is the monitoring terminal equipment with human-computer interaction function, including smart mobile phone, computer
And have the TV of interactive function etc..
Cloud Server is connected with second detection device and monitor terminal by network respectively, for video frequency feature data
Training, according to the information of user feedback, dynamically updates the data sorting technique, wherein,
The video segment that Cloud Server is sent according to monitor terminal extracts characteristic information, and by characteristic information, classification results
Storage parses, and according to renewal sample training data, generate new as new sample to the instruction that monitor terminal is sent
Data classification method configuration file, while send update notification to second detection device.
As shown in Fig. 2 second detection device provided by the present invention includes：Information collecting device and anomalous event judge dress
Put, wherein, as shown in figure 3, the anomalous event judgment means include the first transceiver module, analysis module, data categorization module,
Memory module and the first control module, control module are connected with modules respectively, wherein,
Information collecting device is camera and sound input system；The camera and sound input system by wired or
Wireless networking mode is connected with anomalous event judgment means, and the video information and audio for real-time acquisition monitoring environment are believed
Breath, and the video of collection and audio transmission to anomalous event judgment means are handled.
Analysis module is used to analyze and process the video data of information collecting device collection；
The image information that data categorization module is handled according to analysis module judges whether monitors environment is abnormal.Data classification mould
Block needs the data classification method configuration file (generally use memory-resident or firmware mode) of built-in initialization.Initialization
Data classification method configuration file is obtained by a large amount of mark post sample datas.Data classification method can be understood as a mapping
Relation, can be automatically+1 or -1 by the maps feature vectors of input.Represent that " someone invades respectively with+1, -1 and 0 in the present invention
Critical event ", " the insignificant event for having sound " and the class recognition result of event without exception three.Data categorization module, which identifies, works as
After which kind of preceding video belongs to, recognition result is sent to the monitor terminal of user.Pass through the continuous feedback data of user, cloud service
Device training generates new data classification method.The configuration file that this method is used in the data categorization module of renewal.Such data
The data classification method that sort module can constantly update.And new data classification method is adaptive current monitor environment
, therefore classification is identified using new sorting technique, it can effectively reduce rate of false alarm.
Memory module is used for the video data for storing collection, is easy to checking and playing back for user.Memory module can use
Flash Memory, DDR SDRAM etc. are realized.
Transceiver module is used between anomalous event judgment means and information collecting device, monitor terminal and automatic alarm system
Information exchange；
Control module is the corn module of the anomalous event judgment means, controls the number between modules respectively
According to communication and interaction.The control module can be realized by single-chip microcomputer or microcontroller (MCU) etc., on the one hand receive monitor terminal
Video check instruction after, notice memory module by video data transmission to monitor terminal, on the other hand by under communication module
Carry and be used for the configuration file for updating the data data classification method in sort module in Cloud Server.
As shown in figure 4, analysis module is used to analyze and process the video data of collection, the acquisition being linked in sequence is specifically included
Three module, decomposing module and feature extraction units.These units can be realized in a manner of software or firmware.Wherein, obtain
Module can use frame difference method, optical flow method and dynamic self-adapting background method scheduling algorithm, enabling distinguish background and motion
Object.After the object that acquisition module extraction detects, target is split by decomposing module.Wherein Target Segmentation can be adopted
With Otsu methods (maximum variance between clusters), iterative method, maximum entropy method (MEM) etc..Feature extraction module by the target of segmentation is carried out with
Track, extract the characteristic information of image, such as the feature such as color, shape, movement locus.Wherein image feature information is generally with feature
Vector form is preserved.
As shown in figure 5, the first transceiver module includes video access unit, data interaction unit.Wherein video access unit
For accessing the video of collection, and video is sent into analysis module and memory module.Wherein video access unit is video
Decoding circuit, the video data for video acquisition device collection are decoded.On the one hand data interaction unit is used for abnormal thing
Data interaction between part judgment means and monitor terminal, the classification results for sorting data into module are sent to monitor terminal, together
When receive monitor terminal send video check information；On the other hand it is used between anomalous event judgment means and Cloud Server
Information exchange.Cloud Server is updated the data in sort module by training video characteristic, notice anomalous event judgment means
Classification configurations file.Data interaction unit can be using any one in wired network communication or wireless communication mode
Kind, such as cable network can be using Ethernet interface etc., wireless network communication can be WIFI and 3G/4G etc..
The second detection device also includes：Pattern recognition device, as shown in fig. 6, described image identification device includes：Place
Module, image collection module, matching module, database, identification module, the second transceiver module composition are managed, wherein,
Processing module, for the audio/video information of processing information acquisition module collection, extract the image for face recognition；
Image collection module, for obtaining the image information for face recognition, the digital information of described image is carried out
Analysis and processing, the pretreatment of advanced row digital information, remove mixed interference information and reduce some deformations and distortion, then
Feature extraction is carried out, the facial image information in digital information is obtained with mode identification method；
Matching module, whether the facial image information for judging acquired has storage in database, for being obtained
Facial image information, the content stored in searching data storehouse, be determined as when finding matched information without exception
Situation, send and start door opening command, sentenced when the content matched with the facial information is not present in database by anomalous event
Disconnected device judges whether it is anomalous event；
Database, for storing face information and the audio-frequency information that user pre-sets；
Identification module, for carrying out match cognization, the identification module bag according to existing face information and audio-frequency information
Include face image identification unit and acoustic recognition unit；
Second transceiver module, for realizing second detection device and the communication of monitor terminal and automatic alarm system.
System carries out pattern recognition process, pattern recognition module bag by the corresponding processing method of invocation pattern identification device
Monitoring objective face-image recognition unit and monitoring objective disk acoustic recognition unit are included, wherein monitoring objective face-image identification is single
Acquisition, pretreatment, the extracting method of Invariance feature and neural network recognization technology of the member to target image initial data, no
The extraction of Vertic features extracts one group of feature with consistency after digitlization or in pretreated input pattern；Monitoring
Target sound recognition unit carries out match cognization come all for voice digital signal with corresponding processing method and discrimination method, if
Having the target information, then the match is successful, otherwise judges whether it is anomalous event by anomalous event judgment means.
As shown in fig. 7, the 3rd detection means includes：Video information acquisition module, detection module, control module and Special Graphs
As monitoring module, wherein,
Video information acquisition module is used for continuous acquisition video image, using the video image currently collected as current defeated
Enter image to send to detection module；
Detection module is used to receive the current input image from video information acquisition module, obtains current input image
Background difference image and inter-frame difference image, by the way that the background difference image and the inter-frame difference image are carried out into logical AND
Processing obtains moving image, exports the moving image of acquisition.
Control module is used to receive the moving image for carrying out self-detection module, and video prison is carried out according to the moving image received
Control；
Particular image monitoring module is used to, by merging night interesting target and background image between daytime, obtain final night
Activity detection image.
As shown in figure 8, detection module includes：Motion detection block 01, target tracking module 02, subsequent analysis module 03,
Alarm module 04, video encoding module 05.
Motion detection block 01 is used to receive the current input image from video acquisition device 901, obtains current input
The background difference image of image and inter-frame difference image, by the way that the background difference image and the inter-frame difference image are carried out
Logical AND processing obtains moving image, exports the moving image of acquisition to target tracking module 02.
As shown in figure 9, motion detection block 01 includes：Difference block 801, extraction module 802, the 3rd filtration module
803 and context update module 804.
Difference block 801 is used to receive current input image, is obtained and carried on the back according to current input image and current background image
Scape difference image, inter-frame difference is obtained according to the former frame input picture of the current input image and the current input image
Image, the background difference image of acquisition and inter-frame difference image are sent to extraction module 802.
Wherein, as shown in Figure 10, difference block 801 includes background difference block 8011 and inter-frame difference module 8012.
Background difference block 8011 includes：Background storage module 11, the first subtraction module 12, the and of the first binarization block 13
First filtration module 14.
Background storage module 11 is used to store current background image, and the current background image of storage is output into first subtracts each other
Module.
First subtraction module 12 is used to receive current input image, by the current input image received and background storage mould
Current background image in block 11 subtracts each other, and the background difference image obtained after subtracting each other is exported to first binarization block
13。
First binarization block 13 is used to receive the background difference image from first subtraction module 12, will receive
Background difference image carry out binary conversion treatment, export the background difference image after binary conversion treatment to the first filtration module 14.
First filtration module 14 is used to receive the background difference image after the binary conversion treatment of the first binarization block 13,
Morphologic filtering processing is carried out to the background difference image received, the background difference image after output morphologic filtering processing is extremely
Extraction module 802.
Inter-frame difference module 8012 includes：Postponement module 21, the second subtraction module 22, the second binarization block 23 and second
Filtration module 24.
Postponement module 21 is used to receive input picture, by the way that the input picture received progress delay disposal is obtained currently
The former frame input picture of input picture, the former frame input picture of the current input image is sent to the second subtraction module
22。
As can be seen here, the effect of Postponement module 21 is so that the former frame of current input image and current input image
Input picture can be sent to the second subtraction module 22 simultaneously.
As an example it is assumed that input picture a is input picture a ' former frame input picture, in no Postponement module 21
In the case of, input picture a and input picture a ' will be successively sent to the second subtraction module 22 at T-1 moment and T moment respectively；Such as
Fruit is by input picture by being retransmited after deferred mount 21 to the second subtraction module 22, then should be sent at the T-1 moment to
The a of two subtraction modules 22 will could be sent at the T moment to the second subtraction module 22.Therefore deduce that, at the T moment, second
Subtraction module 22 will receive input picture a ' that video acquisition device 901 directly transmits simultaneously and from Postponement module
21 input picture a.
Second subtraction module 22 is used for before receiving current input image and current input image from Postponement module 21
One frame input picture, the former frame input picture of the current input image received and the current input image is subtracted each other, will
The error image obtained after subtracting each other is exported to the second binarization block 23.
Second binarization block 23 is used to receive the error image from the second subtraction module 22, the differential chart that will be received
As carrying out binary conversion treatment, the inter-frame difference image obtained after output binary conversion treatment to the second filtration module 24.
Second filtration module 24 is used to receive the inter-frame difference image after the binary conversion treatment of the second binarization block 23,
Morphologic filtering processing is carried out to the inter-frame difference image received, the image after output morphologic filtering processing to extraction
Module 802.
Extraction module 802 is used to receive background difference image and inter-frame difference image from difference block 801, will
The background difference image and the inter-frame difference image carry out logical AND processing, the motion diagram obtained after output logical AND processing
As to the 3rd filtration module 803.
3rd filtration module 803 is used to receive the moving image from extraction module 802, to the motion diagram received
Picture carries out morphologic filtering processing, and the moving image after output morphologic filtering processing to central controller 903 and background are more
New module 804.
Context update module 804 is used to receive current input image and the 3rd filtering mould from video acquisition device 901
The moving image that block 803 currently exports, current background is updated according to current input image and the moving image currently exported
Image, the renewal current background image refers to, the current back of the body is updated in the method for testing motion provided according to embodiments of the present invention
The method of scape image, it is current in the moving image renewal background storage module 11 according to current input image and currently exported
Background image.
Target tracking module 02 is used for the moving image for receiving the output of motion detection block 01, in the motion continuously received
Moving target is identified in image, and records the movement locus of the moving target, exports the image after target following to follow-up
Analysis module 03.
Target tracking module 02 can use target in method for testing motion provided in an embodiment of the present invention in the present embodiment
The method of tracking carries out target following.
Subsequent analysis module 03 is used to receive the image from target tracking module 02, the image received is carried out follow-up
Analysis, when analysis result be situation exception, notice alarm module 04 is alarmed, and exports the image that receives to Video coding
Module 05.
Alarm module 04 is used to receive the notice from subsequent analysis module, receives the notice from subsequent analysis module
After alarmed.
Video encoding module 05 is used to receive the image from subsequent analysis module, and the image received is compiled by video
Code method is encoded, and the image after coding is sent to central controller 903 by IP network.
The method for video coding that video encoding module 05 can use has many kinds, such as Mepg4, H.264 waits coding staff
Formula.
Control module is used to receive the image from video encoding module 05, and video prison is carried out according to the image received
Control.
In above-described intelligent monitor system, detection module can use embedded chip to realize, be placed on control module
Front end.
The nighttime image sequence that particular image monitoring module collects to camera first carries out image enhaucament, is then increasing
Moving object detection is carried out on image sequence after strong, obtains night movement target.The illumination of nighttime image sequence is extracted simultaneously
Region.Carried on the back finally by fusion night interesting target (night movement target, the light area of nighttime image sequence) between daytime
Scape image, to obtain final result image.
As shown in figure 11, the particular image monitoring module based on information fusion includes, image enhancement module, target detection mould
Block, region extraction module, background segment module, Fusion Module, it is specially：S1：Image enhancement module collects to camera
Nighttime image sequence carries out image enhaucament, the image sequence higher for obtaining contrast；S2：After module of target detection is to enhancing
Image sequence carry out target detection, for obtaining night movement target；S3：Region extraction module extraction nighttime image sequence
Light area；S4：Background segment module carries out background segment to the image sequence under Same Scene daytime, obtains between clearly daytime
Background image；S5：Night interesting target is night movement target, the light area of nighttime image sequence and daytime by Fusion Module
Between background image carry out image co-registration.
According to an embodiment of the invention, described image enhancing specifically includes：
Step S11：Each two field picture is extracted in the nighttime image sequence collected from camera；
Step S12：The gray value of all pixels in each two field picture is entered into line translation according to a greyscale transformation function, used
In obtaining the higher image sequence of contrast.
According to an embodiment of the invention, the target detection specifically includes：
Step S21：Background model is built to the enhanced image sequence obtained in step S1；
Step S22：Region of variation is extracted from background model from enhanced image sequence, for obtaining night
Between moving target；
Step S23：Obtained night movement target is subjected to morphological operation and connected domain analysis, it is accurate to obtain segmentation
Night movement target.
According to an embodiment of the invention, the extraction of the light area comprises the following steps：
Nighttime image sequence is subjected to gaussian filtering, light area is considered as to the LPF result of image.
According to an embodiment of the invention, the background segment comprises the following steps：
Background model is built the image sequence under Same Scene daytime, obtains clearly background image.
According to an embodiment of the invention, described image fusion comprises the following steps：
Step S51：According to the weight coefficient of night movement target information and night light area information acquisition image co-registration；
Step S52：Night interesting target is night movement target, night light area and background image is carried out between daytime
Weighted Fusion, obtain more comprehensively, clearly describing scene.
In monitoring, moving object is the emphasis of monitoring, and the purpose of target detection is by region of variation from image sequence
Extracted from background image.Mixed Gauss model method is high to the adaptivity of background, and change to brightness, thing in background
The trickle movement of body, at a slow speed target etc. have good adaptability, in the present invention, using mixed Gauss model method to enhancing
Video sequence afterwards carries out background modeling.Because night video sequence signal to noise ratio is low, noise is stronger, mixed Gauss model method institute
Some noise points and cavity are included in obtained foreground image.The present invention removes noise using medium filtering, in addition using shape
The corrosion expansive working of state removes the cavity in foreground image, and each moving object is obtained finally by 8 connected domain analysis
Contours segmentation result.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited thereto,
Any one skilled in the art the invention discloses technical scope in, the change or replacement that can readily occur in,
It should all be included within the scope of the present invention.Therefore, protection scope of the present invention should the protection model with claim
Enclose and be defined.
Claims (8)
1. a kind of intelligent hierarchical monitoring system, including master controller, detection means, automatic alarm system and monitor terminal, its
In,
The detection means includes the first detection means, second detection device and the 3rd detection means, and the master controller is used for
Realize the control and communication to first detection means, second detection device, the 3rd detection means, automatic alarm system；
First detection means is used to carry out primary detection, and it is abnormal living to judge whether each entrance that can lead to interior has
It is dynamic, first detection means is persistently detected if abnormal movement is not detected, otherwise, automatic alarm system is entered
Enter and treat alarm status, and start second detection device and detected；
The second detection device is for detecting whether be anomalous event, if anomalous event, then automatic alarm system is carried out
Alarm, otherwise, start the 3rd detection means and detected；
Whether someone is movable for detecting interior for 3rd detection means, if nobody is movable, repeated priming first is examined
Survey device to be detected, if someone is movable, automatic alarm system enters no alarm state；
Characterized in that,
Automatic alarm system includes voice module and Cloud Server, when automatic alarm system enters alarm status, master controller
Control voice module sends voice warning, while warning message is sent into monitor terminal by Cloud Server, can set each
The different alarm sound of kind；
Monitor terminal is connected with second, third detection means and automatic alarm system by network respectively, for user to prison
Control situation checks that monitor terminal receives the video information and recognition result of warning message and the transmission of second, third detection means
Information, and be determined as that wrong video feature information is sent into Cloud Server by what second detection device was sent through terminal and carry out
Training；
Cloud Server is connected with second detection device and monitor terminal by network respectively, the instruction for video frequency feature data
Practice, according to the information of user feedback, dynamically update the data sorting technique, wherein,
The video segment that Cloud Server is sent according to monitor terminal extracts characteristic information, and characteristic information, classification results are stored
As new sample, the instruction that monitor terminal is sent is parsed, and according to renewal sample training data, generates new data
Sorting technique configuration file, while send update notification to second detection device.
2. a kind of the system as claimed in claim 1, the first detection means includes detecting module, is passed in and out for testing staff each
The activity of entrance, door status sensor, infrared detector, photo-electric blocking sense can be used installed in the first detection means of entry door
Device, microwave sensor or dual technology detector are answered, for detecting the personnel activity of disengaging entry door；Discrepancy beside window
Mouth detection means, double curtain infrared detectors can be used, people can be identified according to the triggering priority time of its two infrared probes
Member enters outgoing direction, tells personnel and stretches out one's hand pass from the abnormal movement got in outside window, daily routines indoors and from interior
Window activity；The first detection means of installation on balcony, can be used infrared detector or dual technology detector, exists for detecting personnel
The activity of balcony.
3. a kind of system as claimed in claim 2, second detection device include：Information collecting device and anomalous event judge dress
Put, the anomalous event judgment means include the first transceiver module, analysis module, data categorization module, memory module, Yi Ji
One control module, control module are connected with modules respectively, wherein,
Information collecting device is camera and sound input system；
Analysis module is used to analyze and process the video data of information collecting device collection；
The image information that data categorization module is handled according to analysis module judges whether monitors environment is abnormal, the data classification mould
Original data classification method configuration information is stored with block, after data categorization module identifies which kind of current video belongs to,
Recognition result is sent to the monitor terminal of user, by the continuous feedback data of user, Cloud Server training generates new data
Sorting technique, the configuration file that methods described is used in the data categorization module of renewal；
Memory module is used for the video data for storing collection, is easy to checking and playing back for user；
The letter that transceiver module is used between anomalous event judgment means and information collecting device, monitor terminal and automatic alarm system
Breath interaction；
First control module is the corn module of the anomalous event judgment means, controls the number between modules respectively
According to communication and interaction.
4. a kind of system as claimed in claim 3, the analysis module includes Target Acquisition module, the decomposition mould being linked in sequence
Block and feature extraction module, wherein,
The acquisition module, for using corresponding algorithm, the object of background and motion is distinguished, then extracts the mesh detected
Mark；
The decomposing module, for after the acquisition module extracts the target detected, splitting to target；The feature
Abstraction module, for by being tracked to the target of segmentation, extracting the characteristic information of image, by described image characteristic information with
Characteristic vector form is preserved.
5. a kind of system as claimed in claim 4, first transceiver module includes video access unit, data interaction list
Member, wherein video access unit are used for the video for accessing collection, and video is sent into analysis module and memory module, its
In,
Video access unit is video decoding circuit, and the video data for information collecting device to be gathered is decoded；Data
The data interaction that interactive unit is used between anomalous event judgment means and monitor terminal, sort data into the classification results of module
Send to monitor terminal, while the video for receiving monitor terminal transmission checks information；It is also used for anomalous event judgment means and cloud
Information exchange between server, Cloud Server pass through training video characteristic, notice anomalous event judgment means renewal number
According to the classification configurations file in sort module, data interaction unit is using in wired network communication or wireless communication mode
Any one, cable network can use Ethernet interface, and wireless network communication can be WIFI and 3G/4G.
6. a kind of system as claimed in claim 5, the second detection device also include：Pattern recognition device, described image
Identification device includes：Processing module, image collection module, matching module, database, identification module, the second transceiver module, its
In,
Processing module, for the audio/video information of processing information acquisition module collection, extract the image for face recognition；Image
Acquisition module, for obtaining the image information for face recognition, the digital information of described image is analyzed and handled, so
After carry out feature extraction, with mode identification method obtain digital information in facial image information；
Matching module, whether the facial image information for judging acquired has storage in database, for the face obtained
Portion's image information, the content stored in searching data storehouse, is determined as situation without exception when finding matched information,
Send and start door opening command, when the content matched with the facial information is not present in database by anomalous event judgment means
Judge whether it is anomalous event；
Database, for storing face information and the audio-frequency information that user pre-sets；
Identification module, for carrying out match cognization according to existing face information and audio-frequency information, the identification module includes face
Portion's image recognition and acoustic recognition unit；Second transceiver module, for realizing second detection device and monitor terminal and automatic report
The communication of alert system.
7. a kind of system as claimed in claim 6, acquisition of the face-image recognition unit to target image initial data,
Pretreatment and the extracting method and neural network recognization technology using Invariance feature；Acoustic recognition unit is directed to voice number
Word signal is all to carry out match cognization with corresponding processing method and discrimination method, and the match is successful if having the target information, no
Then judge whether it is anomalous event by anomalous event judgment means.
8. a kind of system as claimed in claim 7, the 3rd detection means include：Video information acquisition module, detection module, control
Molding block and particular image monitoring module, wherein,
Video information acquisition module is used to gather video image, and the video image currently collected is sent to detection module；
Detection module is used to receive the input picture from video information acquisition module, exports the moving image of acquisition to controlling mould
Block；
Control module is used to receive the moving image for carrying out self-detection module, and video monitoring is carried out according to the moving image received；
Particular image monitoring module is used to, by merging night interesting target and background image between daytime, obtain final nocturnalism
Detection image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610744200.7A CN106327738B (en) | 2016-08-26 | 2016-08-26 | A kind of intelligent hierarchical monitoring system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610744200.7A CN106327738B (en) | 2016-08-26 | 2016-08-26 | A kind of intelligent hierarchical monitoring system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106327738A CN106327738A (en) | 2017-01-11 |
| CN106327738B true CN106327738B (en) | 2017-11-24 |
Family
ID=57791423
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610744200.7A Active CN106327738B (en) | 2016-08-26 | 2016-08-26 | A kind of intelligent hierarchical monitoring system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106327738B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107845105B (en) * | 2017-10-24 | 2021-09-10 | 深圳市圆周率软件科技有限责任公司 | Monitoring method based on panoramic gun-ball linkage, intelligent device and storage medium |
| CN107945442A (en) * | 2017-10-30 | 2018-04-20 | 深圳辉锐天眼科技有限公司 | Alarm management method |
| CN108133178B (en) * | 2017-12-08 | 2021-11-30 | 重庆广睿达科技有限公司 | Intelligent environment monitoring system and method based on image recognition |
| CN109671001A (en) * | 2018-12-11 | 2019-04-23 | 上海东冠通信建设有限公司 | Intelligence community safety monitoring system and method for safety monitoring |
| CN111860030A (en) * | 2019-04-24 | 2020-10-30 | 杭州海康威视数字技术股份有限公司 | Behavior detection method, behavior detection device, behavior detection equipment and storage medium |
| CN112422885A (en) * | 2019-08-22 | 2021-02-26 | 北京奇虎科技有限公司 | Security detection method and device and cloud server |
| CN110807444A (en) * | 2019-12-06 | 2020-02-18 | 太原师范学院 | Pedestrian abnormal behavior detection system and method based on DSP edge calculation |
| CN112001817B (en) * | 2020-08-24 | 2021-04-16 | 山西辰涵数字科技股份有限公司 | Scenic spot real-time guide explanation management system based on artificial intelligence |
| CN113727213A (en) * | 2021-08-26 | 2021-11-30 | 四川华能氢能科技有限公司 | Remote online monitoring system for hydrogen production device and implementation method thereof |
| CN114219400B (en) * | 2021-12-15 | 2022-08-05 | 浙江省邮电工程建设有限公司 | Material supervision system and method of intelligent factory |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4019266B2 (en) * | 2001-10-25 | 2007-12-12 | 日本電気株式会社 | Data transmission method |
| CN101674461A (en) * | 2008-09-11 | 2010-03-17 | 上海市长宁区少年科技指导站 | Intelligent network monitoring system for safety of primary and secondary school campuses |
| CN103198605A (en) * | 2013-03-11 | 2013-07-10 | 成都百威讯科技有限责任公司 | Indoor emergent abnormal event alarm system |
| CN203217687U (en) * | 2013-05-10 | 2013-09-25 | 浙江纳特智能网络工程有限公司 | Intelligent building monitoring system |
| CN105405238B (en) * | 2015-11-27 | 2018-08-17 | 南京邮电大学 | A kind of dormitory safety management system based on infrared technique and wireless sensor technology |
| CN105632058B (en) * | 2016-03-17 | 2018-11-27 | 云丁网络技术(北京)有限公司 | Intelligent alarm method and its alarm system based on human body infrared characteristic identification human body |
-
2016
- 2016-08-26 CN CN201610744200.7A patent/CN106327738B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN106327738A (en) | 2017-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106327738B (en) | A kind of intelligent hierarchical monitoring system | |
| Gong et al. | A Real‐Time Fire Detection Method from Video with Multifeature Fusion | |
| CN103839373B (en) | A kind of unexpected abnormality event Intelligent Recognition alarm device and warning system | |
| CN104573679B (en) | Face identification system based on deep learning under monitoring scene | |
| CN105426820B (en) | More people's anomaly detection methods based on safety monitoring video data | |
| CN105913528A (en) | Method and device for processing access control data, method and device for access control | |
| CN110717389B (en) | Driver fatigue detection method based on generation countermeasure and long-short term memory network | |
| US10445885B1 (en) | Methods and systems for tracking objects in videos and images using a cost matrix | |
| CN109672863A (en) | A kind of construction personnel's safety equipment intelligent monitoring method based on image recognition | |
| KR20160135004A (en) | Module-based intelligent video surveillance system and antitheft method for real-time detection of livestock theft | |
| CN109757395A (en) | A kind of pet behavioral value monitoring system and method | |
| US11354819B2 (en) | Methods for context-aware object tracking | |
| US20220076022A1 (en) | System and method for object tracking using feature-based similarities | |
| US20220301275A1 (en) | System and method for a hybrid approach for object tracking across frames. | |
| CN110490126B (en) | Safe deposit box safety control system based on artificial intelligence | |
| CN116563797B (en) | Monitoring management system for intelligent campus | |
| CN111754669A (en) | College student management system based on face recognition technology | |
| Xu et al. | Unusual event detection in crowded scenes using bag of LBPs in spatio-temporal patches | |
| CN105227918B (en) | A kind of intelligent control method | |
| CN116798176A (en) | Data management system based on big data and intelligent security | |
| CN112866654A (en) | Intelligent video monitoring system | |
| US11163097B1 (en) | Detection and correction of optical filter position in a camera device | |
| CN105139503A (en) | Lip moving mouth shape recognition access control system and recognition method | |
| CN106341658A (en) | Intelligent city security state monitoring system | |
| CN109300279A (en) | A kind of shop security monitoring method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |